Redis Cachify

Simple Redis & Memory Caching Library

Install

npm install redis-cachify --save

Usage

This library requires either existing redis connetions, or redis host and port settings.

e.g.

Cachify = require("redis-cachify")
cachify = Cachify
  redisPort: 6379
  redisHost: "localhost"

# OR

cachify = Cachify
  redisStore: redisClient (in normal mode)
  redisPubsub: redisClient (in pubsub mode)

# Other options (with defaults are):

cachify = Cachify
  redisPort: 6379
  redisHost: "localhost"
  defaultExpiry: 60 # (time in seconds for functions to be cached)
  cacheInProcess: false # (turns on in-process cachcing, see details below)
  redisChannel: "cache" # (the Redis pubsub channel to use)

Once you have a cachify instance you can use it as follows:

cachifiedFunction = cachify(id, myFunction)

This library currently expects all functions that it as applied to, to have the following signature:

myFunction = (data, callback) ->

Callbacks are standard Node-style callbacks that are expected to receive an optional error and then a result.

id can be a string, e.g. - "usd2gbp' or a function, e.g. (data) -> "userToken:" + data.userId

If the id argument is a function, then it will be passed the data argument passed in each time the function is called.

A cachified version of a function can be used in excactly the same way as the function that was passed into it. The difference being that the result will be cached in Redis and returned much sooner. This can be very useful in the following scenario:

Promises

This library can now be used with functions that return promises as follows:

cachifiedFunction = cachify.promise(id, myFunction)

where:

myFuncion = (data) -> new Promise ...

Getting a third party auth toekn that lasts for n minutes

Application logic is vastly simplified as all you need to do is "cachify" the function that goes and gets the token. Any method running in any process (on the same or different machines) can now reliably call that method. The function will return either a fresh token, or a cached token that will automatically expire after n minutes. If 3 methods call the function in the same tick of the event loop, the original function will only be called once - yet the callbacks from all 3 functions will correctly be called

If you make a 10 calls to cachifiedFunction (with the same data), then myFunction will only be called once.

Custom Expiry per Function

Simply pass in the expity time in seconds as the second argument:

cachifiedFunction = cachify(id, 120, myFunction)

In Process Cache

If inProcessCache is set to true then the result of myFunction will be saved both in Redis and in a local memory cache. This can improve performance by preventing unncecesary IO to Redis, but could result in slight race conditions - so it is turned off by default. The race conditions would only occur at the time of a results expiry from the cache. The in process cache could be slighly slower at expiring the data than Redis. This is highly unlikely to cause a problem - but you should be aware of it.

Troubleshooting

This library has a basic test suite and has been used extensively in production. If you are having problems, try running your program with the environment variable DEBUG=cachify This will print a whole load of debugging information to the console.

Todo

• Enforce Redis 2.6 and use new set method signature with expiry
• More tests for in process caching